Chinese AI labs like Moonshot AI and Alibaba are releasing powerful open-source models, challenging Western dominance. US officials, including Commerce Secretary Scott Bessent and Michael Kratsios, express concern, alleging IP theft and considering sanctions, intensifying the US-China AI competition.
Chinese start-up Moonshot AI has launched K3, an open-source AI model claimed to be the world's largest by parameters, asserting it outperforms Anthropic's Claude Fable 5 and OpenAI's GPT-5.6 Sol in specific benchmarks.
Moonshot AI launched Kimi K3, a 2.8-trillion-parameter open-source AI model, outperforming Claude Fable 5 in scriptwriting and frontend coding benchmarks. It offers near-frontier performance at mid-tier prices, demonstrating Chinese innovation despite US chip export controls.
Chinese President Xi Jinping used the World AI Conference in Shanghai to promote open-source AI, challenging US dominance and advocating for global collaboration. China launched WAICO, a 29-nation coalition, to give Global South nations a voice in AI governance, while also developing its own advanced open-weight models.
MiniMax CEO Yan Junjie announced he will not take a salary until the company achieves artificial general intelligence (AGI). He also pledged to transfer 4% of his personal shares to employees and 1% to support open-source AI development. The company is also seeking to raise approximately HK$16.04 billion.
The AI race between the US and China is shifting to a competition over electricity costs and output, as AI models become commoditized. China's focus on renewable energy, particularly solar power, offers a path to cheaper electricity for data centers, potentially challenging US dominance.
DeepReinforce launched Ornith-1.0, a family of open-source AI coding models available in four sizes (9B to 397B parameters) under MIT license. Optimized for agentic coding tasks, these models can autonomously perform multi-step development workflows, outperforming competitors like Claude Opus on specific benchmarks.
Anthropic's recent suspension of AI models due to a US government directive underscores the risks of vendor lock-in. Companies are increasingly seeking downloadable, open-source models to maintain control over their intellectual property and avoid disruptions, especially as Chinese AI labs gain traction.
Meta's Chief AI Officer Alexandr Wang stated their open-source AI strategy is unsuitable for frontier models like Muse Spark due to bio and safety risks. The company is exploring subscription models for its platforms and AI chatbot to diversify revenue beyond advertising.
Meta's Chief AI Officer Alexandr Wang admits the company's open-source AI strategy is outdated for frontier models like Muse Spark, which was kept proprietary due to bio and safety risks. Meta is exploring subscription models for its platforms and AI chatbot to diversify revenue beyond advertising.
Meta's Chief AI Officer Alexandr Wang revealed that the company's open-source AI strategy is no longer suitable for frontier models like Muse Spark, which was kept proprietary due to bio and safety risks. Rival labs face similar issues as models scale. Meta is also exploring subscription models for its platforms and AI chatbot to diversify revenue beyond advertising.
Zoho founder Sridhar Vembu advised Indian organizations to focus on open-source AI alternatives after the US restricted access to Anthropic's advanced models, highlighting technology's link to national security and sovereignty.
Elon Musk is set to go to trial against Sam Altman's OpenAI on April 27, 2026, in a lawsuit marking the culmination of a years-long feud over AI development. Musk, a co-founder who contributed roughly $100 million in 2015, claims OpenAI betrayed its nonprofit mission to develop AI for humanity's benefit, transforming into a closed-source for-profit subsidiary of Microsoft, which invested $13 billion. OpenAI dismisses the lawsuit as distraction driven by competitive jealousy, noting Musk now competes through his own xAI venture.
DeepSeek released its V4 foundational AI model in two versions - V4-pro with 1.6 trillion parameters and V4-flash with 284 billion parameters. Both feature a 1 million token context window, up from 128,000 in the previous model. The open-source models aim to compete with US leaders OpenAI and Google DeepMind, with Huawei and Cambricon quickly announcing chip compatibility support.
DeepSeek has released its next-generation V4 open-source AI model in two versions: V4-pro with 1.6 trillion parameters and V4-flash with 284 billion parameters. Both models feature a 1 million token context window, achieved with 'world-leading' cost efficiency, positioning the Hangzhou startup as competitive with leading US closed-source models.
In China’s hyper-competitive and lucrative tech industry, releasing cutting-edge artificial intelligence models for free may seem counter-intuitive – but it has become a core business strategy. At the University of Hong Kong last November, Alibaba Group Holding chairman Joe Tsai was asked why the tech giant open-sourced its AI models. Addressing a room full of students, Tsai said he believed open-source AI would bring global benefits by lowering costs, making it a natural fit for talent- and...
